Should Engine healthchecks rely on Appetiser?

[donaldgray]

Engine has a couple of configured health-checks:

• Database (via AddDbContextCheck)
• Queue listeners - this is a custom health-check that validates all configured listeners are still running and haven't bombed out.

Should we add Appetiser as a health-check for Engine? During batch ingest for SLF there were a couple if instances were Appetiser appeared to crash - it should restart of it's own accord but it might be easier for this failure to mark Engine as a whole as faulty. Alternatively stop listening to Image queue until Appetiser is healthy again - potentially reject sync ingest requests during this window too.